Update app.py

app.py

@@ -6,7 +6,7 @@ import pandas as pd
 from transformers import pipeline
 from huggingface_hub import hf_hub_download
 
-# **📌 先运行 `preprocess.py`，确保数据可用**
+# **📌 先运行 `preprocess.py` 处理数据**
 processed_data_path = "processed_data.csv"
 
 if not os.path.exists(processed_data_path):
@@ -28,11 +28,15 @@ model.load_model(model_path)
 
 # **📌 LLM 解析用户输入**
 def chat_with_llm(user_input):
-    prompt = f"请从以下问题中提取出预测所需的参数（年份、期号、中奖号码等）：'{user_input}'"
+    prompt = f"请仅返回预测所需的参数（年份、期号、中奖号码），格式为：'年份:2025, 期号:16, 号码:[5,12,23,34,45,56], 特别号码:7'。输入问题: {user_input}"
     
     try:
-        response = generator(prompt, max_length=100, num_return_sequences=1)
+        response = generator(prompt, max_length=50, num_return_sequences=1)
         extracted_text = response[0]['generated_text']
+
+        # **🚀 确保 UTF-8 编码，移除特殊字符**
+        extracted_text = extracted_text.encode("utf-8", "ignore").decode("utf-8").strip()
+        
         return extracted_text
     except Exception as e:
         return f"❌ GPT-Neo 处理错误: {str(e)}"
@@ -55,16 +59,26 @@ def predict_lottery(year, period, num1, num2, num3, num4, num5, num6, special):
 
     # **进行预测**
     prediction = model.predict(features)
-    return prediction
+
+    # **🚀 修正浮点数问题：四舍五入为整数**
+    prediction = np.round(prediction).astype(int)
+
+    return prediction.tolist()
 
 # **📌 结合 LLM 和 XGBoost**
 def predict_and_interact(user_input):
     llm_output = chat_with_llm(user_input)
 
-    # **默认解析出的预测参数**
-    year, period = 2025, 16
-    nums = [5, 12, 23, 34, 45, 56]
-    special = 7
+    try:
+        # **解析 LLM 输出格式："年份:2025, 期号:16, 号码:[5,12,23,34,45,56], 特别号码:7"**
+        parts = llm_output.split(",")
+        year = int(parts[0].split(":")[1])
+        period = int(parts[1].split(":")[1])
+        nums = [int(x) for x in parts[2].split(":")[1].strip("[]").split()]
+        special = int(parts[3].split(":")[1])
+
+    except Exception as e:
+        return f"❌ LLM 解析数据失败: {str(e)}\n\n📢 原始 LLM 解析结果: {llm_output}"
 
     # **进行预测**
     prediction = predict_lottery(year, period, *nums, special)